Add prediction endpoint

StreamServer/src/config.py

@@ -28,4 +28,7 @@ CONFIG_FILE = config('CONFIG_FILE')
 YOLO_WEIGHT_FILE = config('YOLO_WEIGHT_FILE')
 SPPE_WEIGHT_FILE = config('SPPE_WEIGHT_FILE')
 TSSTG_WEIGHT_FILE = config('TSSTG_WEIGHT_FILE')
-LINE_NOTIFY_TOKEN = config('LINE_NOTIFY_TOKEN')
+LINE_NOTIFY_TOKEN = config('LINE_NOTIFY_TOKEN')
+WEATHER_API_KEY = config('WEATHER_API_KEY')
+LAT = config('LAT')
+LON = config('LON')

StreamServer/src/query/prediction.py

@@ -0,0 +1,33 @@
+from datetime import datetime, timedelta
+
+from sqlalchemy import func
+from sqlalchemy.orm import Session
+
+from models import PredictionData
+
+
+def get_temp_prediction_data(session: Session):
+    """Get indoor temperature prediction for the next five days"""
+    current_time = datetime.now()
+    limit_day = current_time + timedelta(days=5)
+
+    return session.query(
+        PredictionData.timestamp, PredictionData.indoor_temp
+    ).filter(
+        PredictionData.timestamp >= current_time,
+        PredictionData.timestamp < limit_day
+    ).order_by(PredictionData.timestamp.desc()).all()
+
+
+def get_feature_prediction_data(session: Session):
+    """Get all features used to predict indoor temperature for the next five days"""
+    current_time = datetime.now()
+    limit_day = current_time + timedelta(days=5)
+
+    return session.query(
+        PredictionData.timestamp, PredictionData.outdoor_temp, PredictionData.outdoor_feels_like,
+        PredictionData.outdoor_pressure, PredictionData.outdoor_humidity
+    ).filter(
+        PredictionData.timestamp >= current_time,
+        PredictionData.timestamp < limit_day
+    ).order_by(PredictionData.timestamp.desc()).all()

StreamServer/src/routers/prediction.py

@@ -0,0 +1,78 @@
+import httpx
+
+from datetime import datetime
+
+from sqlalchemy import func
+from sqlalchemy.orm import Session
+
+from fastapi import APIRouter, Depends, HTTPException, status
+
+from analytic.health.indoor_model import XgboostIndoorModel
+from config import WEATHER_API_KEY, LAT, LON
+from database import SessionLocal
+from scheme import IndoorTemperature
+from models import PredictionData
+from query.prediction import get_temp_prediction_data, get_feature_prediction_data
+
+router = APIRouter()
+
+
+#Dependency
+def get_db():
+    db = SessionLocal()
+    try : 
+        yield db
+    finally:
+        db.close()
+
+# Load prediction data -> FOUND -> Return
+# NOT FOUND
+# Try load data from database -> Load future data from database ->  predict -> RETURN
+#                             -> Not found -> Load data from api -> Predict/Save to database -> RETURN
+
+
+def _fetch_data_from_api() -> dict:
+    # api.openweathermap.org/data/2.5/forecast?lat={lat}&lon={lon}&appid={API key}
+    url = f"https://api.openweathermap.org/data/2.5/forecast?lat={LAT}&lon={LON}&units=metric&appid={WEATHER_API_KEY}"
+    response = httpx.get(url)
+    return response.json()
+
+
+@router.get("/indoor/predict/", response_model=list[IndoorTemperature])
+async def get_tomorrow_indoor_temp(db: Session = Depends(get_db)):
+    result = get_temp_prediction_data(db)
+    if not result:
+        features = get_feature_prediction_data(db)
+        xgboost = XgboostIndoorModel()
+        if not features:
+            datas = _fetch_data_from_api()
+            # Save to database
+            if datas:
+                results = []
+                for data in datas['list']:
+                    ts = datetime.fromtimestamp(data['dt'])
+                    data = data['main']
+                    features = [data['temp'], data['feels_like'], data['pressure'], data['humidity']]
+                    result = xgboost.predict(features)
+                    results.append(IndoorTemperature(indoor_temp=result, timestamp=ts))
+                    new_data_entry = PredictionData(
+                        timestamp=ts,
+                        indoor_temp=result,
+                        outdoor_temp=data['temp'],
+                        outdoor_feels_like=data['feels_like'],
+                        outdoor_pressure=data['pressure'],
+                        outdoor_humidity=data['humidity']
+                    )
+                    db.add(new_data_entry)
+                    db.commit()
+                result = results
+            else:
+                raise HTTPException(
+                    status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+                    detail="Failed to fetch data from API"
+                )
+        else:
+            for i in features:
+                indoor = xgboost.predict([i.outdoor_temp, i.outdoor_feels_like, i.outdoor_pressure, i.outdoor_humidity])
+                result.append(IndoorTemperature(timestamp=i.timestamp, indoor_temp=indoor))
+    return result